Re: [SOAF] Please Don't Conflate Direct with Harvested CRs (Central Repositories), Or Deposit Locus With Search Locus

From: Stevan Harnad <amsciforum_at_GMAIL.COM>
Date: Tue, 10 Feb 2009 08:36:03 -0500

       if a central depository is correlated with a political
      entity, particularly a national government, then it may
      be in a position to secure a national mandate in one
      single move.

Indeed. And France's HAL has been correlated with a political entity:
the French government. But in nearly a decade there has not been one
single move toward adopting a nation-wide OA mandate for HAL in

(In the meantime, the UK, with no national repository, has adopted an
OA mandate for all 7 of its national research funding councils , and
7 of its universities have already adopted institutional or
departmental OA mandates too. In France one national funding
council one national institute, and one laboratory have so far
adopted OA mandates.)
      NIH wanting to [manage] its own [r]epository... ha[s] not
      impeded, quite the contrary, the political process
      leading to the law where this mandate is contained.

The issue -- to repeat -- is not institutional repositories (IRs) vs.
central repositories (CRs) but IRs vs. CRs as the mandated locus of
direct deposit. (The rest is indeed just a matter of harvesting.)

And the NIH's stipulation of a CR (PubMed Central) as the mandated
locus of deposit has indeed (partially) impeded (and is continuing to
(partially) impede) progress in implementing and sustaining the NIH

Instead designating the fundee's IR as the default locus of direct
deposit (and then harvesting to PubMed Central from there) would
greatly facilitate the progress not only of the NIH mandate, but of
institutional mandates ("the slumbering giant") as well as other
funder mandates worldwide, for the specific reasons that have been
repeatedly adduced in this discussion.

      I am convinced that we will have to live with a mixed
      array of multi-disciplinary depositories, single (or made
      up of related disciplines) central depositories and
      multi-disciplinary depositories.

To repeat. This is not the issue. Of course there can and will and
should be both IRs and CRs. The issue is locus of mandated deposit. 
      the best way to move forward is simply ensure the
      creation of meta-data that ease[s] the institutional
      harvesting of articles by institutions

(1) Most Institutions do not have OA mandates, hence they are not
harvesting institutional deposit that have been deposited

(2) Most funders do not have OA mandates either (and not all research
is funded), hence much research would not be harvestable to IRs even
if all funders mandated deposit in CRs.

(3) Depositing institution-externally in CRs and then back-harvesting
to the institution's website makes about as much sense as depositing
in google and than back-harvesting to one's own website (except that
it is even worse, with a variety of different googles in which one's
own content might be externally located). Only forward-harvesting --
from the distributed local providers to external collections and
global services, rather than the reverse -- makes any practical or
logical sense.

(4) More important even than making practical and logical sense is
the likelihood that funders mandating IR deposit instead of CR
deposit will facilitate and motivate the adoption of complementary
mandates by the fundees' institutions, to capture the rest of their
institutional research output. (Waking the "slumbering giant": the
universal providers, the worlwide network of universities and
research institutions.) This potential gain would be at no loss in
either content of functionality to the funder.

All of this vague and cheery talk about letting 1000 flowers bloom
simply fails to take into account a single one the many very specific
practical and functional points that have been adduced on behalf of
funders stipulating IRs instead of CRs as their designated locus of
deposit, and harvesting to CRs from there.

Stevan Harnad
Received on Tue Feb 10 2009 - 13:36:50 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:49:41 GMT